Forests and Fuel: Development of a Simple Biomass Comparison Model

Forest-sourced biomass combustion is a popular climate change mitigation technology used to decarbonise electricity generation. Disagreement in the literature on the sustainability of biomass deployment limits policy development. (Chapter 1). The project addresses this uncertainty by development of the Simple Biomass Comparison Model (SBCM) to explore the effect of contrasting assumptions and experimental designs.

An existing model first developed by Sterman et al. (2018a) was identified, analysed in detail and replicated in Python to form SBCM (Chapter 2). SBCM produced a good (but inexact) match for the training data and previously published results. To improve this, the forest growth component of SBCM was re-parameterised (Chapter 3) against the original training data. A significant divergence (p = 0.00002) in species with long growth curves arising from numerical instability in the forest growth function was identified. Analysis of the original supply-chain (Chapter 4) revealed a number of parameterisation errors. These were corrected, and new scenarios for BECCS and gas were developed. These led to large decreases in payback period.

SBCM was modified to improve several inaccurate assumptions in the original model (Chapter 5) by introducing variable rotation length and silvicultural thinning. Shorter rotation lengths resulted in a mean increase in modelled yield of 10.9 GJ per ha per year for non-plantation forests and 7.7 GJ per ha per year for plantations. This highlights the weakness of payback as a suitable metric for biomass uptake.

The study concludes (Chapter 6) that conventional biomass use may be more appropriate than other technologies in some contexts over some time-periods, but that this is by no means certain. Without some form of BECCS technology, biomass remains a low carbon option at best, and is heavily dependent on a sustainable supply chain to achieve positive environmental outcomes.

Further work to develop clear methods and processes is strongly recommended